Two wallets accounted for 99.75% of trading in a Los Angeles election-market episode.
In a Los Angeles mayoral market, two wallets traded directly with each other 1,383 times over about two days in May 2026. Their matched trades totaled $387,979.32 and accounted for 99.75% of all trading we reconstructed in that market during the same period.
Separately, a group of three wallets repeatedly traded with one another in a California congressional primary market and a Texas Senate primary market from December 19, 2025, through January 12, 2026. The public record does not establish who controlled either group or why they traded this way.
Contents
Finding record
Scroll to see all columns
| Field | Record |
|---|---|
| Case ID | MSR-CASE-12 |
| Status | Published September 23, 2026. Research checkpoint: September 10, 2026. |
| Finding | A two-wallet pair generated nearly all reconstructed trading during its roughly 47-hour Los Angeles episode. A separate three-wallet group repeatedly traded with itself in two other election markets. |
| Scope | Three specified Polymarket election markets. The Los Angeles percentage covers the pair's trading window; the California and Texas percentages cover each market's full trading history. |
| Evidence | Public Polymarket trade records and Polygon transactions retained and reconciled by MSR Decode. The figures below are MSR's research results, not published Polymarket estimates. |
| What this establishes | The measured concentration and repeated matched trading, plus the specified on-chain cash outcomes. |
| What it does not establish | Common control of the two groups, intent, wash trading, market manipulation, or any trader's complete account-level profit and loss. |
| Analyst | Chris Park · MSR Decode · ceo@msrdecode.com |
01 · How we counted the trading
On Polymarket, one wallet can buy an outcome from another wallet. We counted a trade between the wallets in this report only when the public record showed them on opposite sides of the same matched transaction. Merely trading in the same market did not qualify.
We counted each matched trade's dollar amount once. For each percentage, we compared the group's matched trading with all trading reconstructed in the stated market and time period using the same accounting method. This measure differs from Polymarket's displayed volume, which counts both sides of a trade. We have not reconciled the two measures.
Scroll to see all columns
| Market and group | Matched transactions | Dollars matched within the group | Share and comparison period |
|---|---|---|---|
| Los Angeles mayoral market · two wallets | 1,383 | $387,979.32 | 99.75% of all reconstructed market trading during the pair's roughly 47-hour episode |
| California 11th district primary · three wallets | 2,247 | $25,387.39 | 27.08% of the market's full trading history |
| Texas Senate second-place market · the same three wallets | 2,376 | $27,186.38 | 25.45% of the market's full trading history |
The percentages answer different questions. The Los Angeles figure covers the two days when the pair traded. The California and Texas figures use each market's entire history because we have not calculated a comparable whole-market total for just the trio's active period. They should not be compared as if they used the same time window.
02 · Nearly all trading during the Los Angeles episode
The two wallets' first matched transaction in the Rae Huang mayoral market was on May 15, 2026. Over the following 47 hours, they traded with each other in 1,383 transactions worth $387,979.32. We reconstructed $388,935.57 of trading by all wallets in that market during the same period. The pair therefore accounted for 99.75% of that activity.
For context, the pair's trades were 82.08% of the market's reconstructed trading across its full history. That is a separate, broader comparison; it does not reduce or replace the two-day figure.
After tracing the wallets' trading and settlement on Polygon, we calculated a combined on-chain cash loss of $71.23 on this market. The exact result before rounding was −$71.228998. This is a measured cash-flow result, not a finding about the wallets' motives.
03 · The same three wallets in California and Texas
A different group of three wallets appeared on both sides of matched trades in a market about Connie Chan in California's 11th congressional district primary and a market about Wesley Hunt placing second in the Texas Republican Senate primary. The group's trading in both markets ran from December 19, 2025, through January 12, 2026. Its first Texas transaction came 148 seconds before its first California transaction.
We counted 2,247 transactions among only these three wallets in California, worth $25,387.39. In Texas, we counted 2,376, worth $27,186.38. Transactions involving an outside wallet were excluded from these two counts.
The California trades made up 27.08% of that market's full reconstructed trading history. The Texas trades made up 25.45% of its full history. Those percentages show substantial repeated trading within the group. They do not show that the three wallets dominated either market throughout its lifetime.
04 · What happened to the money
In Texas, all three wallets had measured cash losses after their trading and settlement records were reconciled: $13.90, $13.69 and $11.27. Using the unrounded amounts, their combined loss was $38.87.
California had a different result. Two wallets closed out with cash losses of $51.04 and $27.49. At the September 10, 2026 research checkpoint, the third wallet still held 39.762124 winning-side shares that it had not redeemed. Those shares had a nominal payout value of $39.76.
Counting that uncollected payout gives the third wallet a marked loss of $20.82. Its cash loss without the payout was $60.58. Neither figure should be described as a final realized result until the remaining position is resolved. This report does not establish whether the wallet redeemed after the checkpoint.
The cash results cover the reconstructed on-chain trading and settlement flows. Our public-record checks found no matching rewards, but they cannot rule out every historical reward or off-chain account adjustment. Accordingly, we do not present these results as complete account statements.
05 · What the pattern does and does not show
The Los Angeles pair and the California–Texas trio are separate observed groups. The trio's activity ended in January; the Los Angeles pair began in May. We found no evidence that the groups traded with each other, shared a controller, or operated as one episode.
Repeated trading between a few wallets can have several explanations. The records establish who traded with whom and how much, but they do not establish why. We therefore make no finding of wash trading, market manipulation, or unlawful conduct.
The transaction counts also use slightly different inclusion rules. The Los Angeles count includes three transactions in which another wallet also participated; only the amount directly matched between the pair is included. The California and Texas counts exclude transactions with any outside participant. These definitions preserve the certified figures, but the counts are not identical measures of participation.
06 · Scope and sources
This is a historical review of three specified markets, not a survey of all Polymarket election trading. The underlying public records are Polymarket's markets and trade data and Polygon's public transaction ledger. MSR retained the transaction-level records and reconciliation behind each figure; wallet addresses and transaction hashes are available on request.
The California wallet's unredeemed position is dated to September 10, 2026. The report does not claim that its status is unchanged today. Nor does it establish whether the same people controlled different wallets. The supported finding is narrower and concrete: two wallets accounted for nearly all trading during one Los Angeles episode, while a separate trio repeatedly traded with itself across two earlier election markets.